Protein kinase C (also known as "calcium/phospholipid-dependent protein kinase", "PKC" or "C-kinase) is a family of very closely related enzymes; one or more members of the protein kinase C family are found in nearly all animal tissues and animal cells that have been examined. The identity of protein kinase C is generally established by its ability to phosphorylate certain proteins when adenosine triphosphate and phospholipid cofactors are present, with greatly reduced activity when these cofactors are absent. Protein kinase C is believed to phosphorylate only serine and/or threonine residues in the proteins that are substrates for protein kinase C. Additionally, some forms of protein kinase C require the presence of calcium ions for maximal activity.
Protein kinase C activity is also substantially stimulated by certain 1,2-sn-diacylglycerols that bind specifically and stoichiometrically to a recognition site or sites on the enzyme. This site is called the diacylglycerol binding site, and it is located on the amino-terminal portion of protein kinase C, the so-called "regulatory domain". The carboxy-terminal portion of protein kinase C carries the site at which protein phosphorylation is effected, and this portion thus called the "kinase domain".
Thus, the rate at which various protein kinase C family members carry out their enzymatic phosphorylation of certain substrates can be markedly enhanced by the presence of the cofactors such as phospholipids, diacylglycerols and, for some protein kinase C family members, calcium ions. This stimulation of protein kinase C activity is referred to as protein kinase C "activation", and the activation of protein kinase C by the binding of diacylglycerols to the regulatory domain of protein kinase C is of particular importance in the normal and pathological functions of protein kinase C.
In contrast to the activation of protein kinase C, some chemical compounds have been shown, when added to protein kinase C enzyme assays, to reduce the rate at which protein kinase C phosphorylates its substrates; such compounds are referred to as protein kinase C "inhibitors" or, in some cases, "antagonists". In some circumstances, protein kinase C inhibitors are capable of inhibiting various cellular or tissue phenomena which are thought to be mediated by protein kinase C.
Activation of protein kinase C by diacylglycerols has been shown to be an important physiological event that mediates the actions of a wide variety of hormones, neurotransmitters, and other biological control factors such as histamine, vasopressin, .alpha.-adrenergic agonists, dopamine agonists, muscarinic cholinergic agonists, platelet activating factor, etc. [see Y. Nishizuka, Nature 308: 693-698 (1984) and Science 225: 1365-1370 (1984) for reviews].
The biological role of protein kinase C is also of great interest because of the discovery that certain very powerful tumor promoting chemicals activate this enzyme by binding specifically and with very high affinity to the diacylglycerol binding site on the enzyme. In addition to diacylglycerols, there are at present six other known classes of compounds that bind to this site: diterpenes such as the phorbol esters; indole alkaloids (indolactams) such as the teleocidins, lyngbyatoxin, and indolactam V; polyacetates such as the aplysiatoxins and oscillatoxins; certain derivatives of diaminobenzyl alcohol; macrocyclic lactones of the bryostatin class; and benzolactams such as (-)-BL-V8-310. The phorbol esters have long been known as powerful tumor promoters, the teleocidins and aplysiatoxins are now known to have this activity, and it appears likely that additional classes of compounds will be found to have the toxic and tumor promoting activities associated with the capability to bind to the diacylglycerol site of protein kinase C and thus activate the enzyme. Other toxicities of these agents when administered to animals include lung injury and profound changes in blood elements, such as leukopenia and neutropenia.
Representative examples of these seven classes of previously known protein kinase C-activating compounds, collectively referred to herein as "phorboids", are depicted below: ##STR1##
It can be seen that the phorboids depicted have diverse structural elements of both hydrophilic and hydrophobic nature, with one prominent exception, namely that each contains a hydroxymethyl or 1-hydroxyethyl group (indicated by the dashed-line boxes in each structure). In each case the phorboid depicted is among the most potent of its particular structural class, and among the seven classes the diterpenes, indolactams, polyacetates, bryostatins and benzolactams have members of especially high potency.
In addition to potent tumor promoting activity, these seven classes of compounds display a vast range of biological activities, as would be expected from the widespread distribution of their target enzyme. Some of these activities, like tumor promotion, indicate the involvement of protein kinase C in important normal or pathological processes in animals. Thus, the phorboids are potent skin inflammatory agents, cause smooth muscle contraction in several tissues, alter immune system function and can be used to cause a variety of other normal or pathological responses. Related disease states such as the development of cancer, the onset and/or maintenance of inflammatory disease, the role of vasoconstriction in hypertension, the role of bronchoconstriction in asthma, the life cycles of many pathogenic human viruses, and the role of cholinergic, adrenergic, and dopaminergic synapses in diseases of the central/peripheral nervous systems, may be mediated in vivo by the stimulation of protein kinase C or other diacylglycerol binding site-bearing entities by diacylglycerols, the latter being generated in the cell by pathological agents or conditions.
In analyzing the activity of a pharmaceutical or other bioactive compound, it is useful to consider two properties: the efficacy, defined as the capability to elicit a full or partial biological result, such as complete displacement of a ligand from its receptor site or the complete inhibition of inflammation or edema caused by a standard stimulus; and the potency, defined as that amount or concentration of drug that causes 50% of the full response (often abbreviated as the ED.sub.50). It is frequently the case within a given class of pharmaceutical agents that individual members of the class all have equal efficacy, i.e. they each can generate a full biological effect, but they show differing potencies. Thus, the structural modifications within such a class affect only the amount necessary to achieve a given result, and the modified compounds otherwise have generally the same central biological characteristic. There may also be differences between members of such a class as regards properties other than the central biological characteristic; for example, members of the class might differ in side effects or toxicity.
Well-known pharmaceuticals that have been in extensive use for years or decades show a wide range of optimal therapeutic potencies. Aspirin, for example, is often taken in multi-gram amounts per day for treatment of inflammation or arthritis, and detailed analyses of its mechanism of action in vitro show that a concentration in the millimolar range is required. In contrast, steroid-based topical anti-inflammatory compounds such as fluocinolone acetonide are many thousand-fold more potent, and, beyond this, some oral contraceptive agents are prescribed in daily doses in the microgram range. Thus, although high potency is generally advantageous for a pharmaceutical, it is not an absolute requirement.
A thousand or more analogs of the highly skin-inflammatory and tumor-promoting phorboids have been reported in the literature, including numerous examples on which minor chemical modifications have been made [see Evans and Soper, Lloydia 41: 193-233 (1978) and references cited therein]. The structures of these phorboids can be compared, and their activities for inflammation and tumor promotion can be analyzed from the perspective of efficacy and potency. The structures of the different classes of phorboids vary quite markedly from one to the other class, yet widespread testing of their biological activities has shown that these classes have generally very similar biological properties. In particular, the numerous known phorboids of the highly potent diterpene, indolactam, and polyacetate classes appear to have, with very minor exceptions, virtually identical efficacies as skin irritants and tumor promoters [T. Sugimura, Gann 73: 499-507 (1982)]. The exceptions involve a few compounds that have a short duration of irritant activity and/or manifest diminished tumor promoting activity, perhaps due to toxicity or secondary parameters such as differing metabolic destruction rates.
In contrast to the essentially equal efficacies among the vast majority of phorboids, their relative potencies cover a wide range, as measured in inflammation and promotion tests and as measured in numerous other in vivo and in vitro systems. Example compounds can be found in the diterpene, indolactam, and polyacetate classes that have nearly equal, very high potencies. At the same time there are compounds in each of these classes which embody significant structural changes that do not diminish efficacy but do result in potency decreases of 10-fold to 100,000-fold or more [see, for example, Driedger and Blumberg, Cancer Res. 37: 3257-3265 (1977), Cancer Res. 39: 714-719 (1979)]. Thus, all these compounds appear to be capable of achieving generally the same biological results, and merely differ in the amount which must be used to obtain a given result.
In vitro measurements of biochemical properties provide an even more sensitive method for comparing the properties of the various phorboids. For example, using a radioactively labeled phorboid such as [.sup.3 H]phorbol 12,13-dibutyrate or [.sup.3 H]lyngbyatoxin, one can measure the potency of a test compound as a competitive ligand for the diacylglycerol binding site, which is also referred to herein as the "phorboid binding site" on protein kinase C or on other biological molecules which have phorboid binding sites (see below). Alternatively, one can measure the ability of a given phorboid to stimulate the protein kinase C-mediated incorporation of radioactive phosphate from [.sup.32 P]adenosine triphosphate into a standard acceptor substrate such as histone H1. Tests of this nature reveal a difference in potency between given phorboid agonists of as much as 10,000,000-fold or more [Dunn and Blumberg, Cancer Res. 43: 4632-4637 (1983), Table 1].
These basic data regarding the phorboid agonists are an important consideration because they underscore the concept that the structural differences among these previously known phorboids, especially the diterpenes, indolactams, polyacetates, and bryostatins, generally do not affect their efficacies as toxic agonists, and indeed a wide variety of structural changes are tolerated in this regard. Such changes generally alter potency only and do not provide agents with therapeutic utility, since the resulting compounds retain their toxicity.
Some minor changes in phorboid structure are known to result in generally inactive compounds, such as a stereochemical change from 4-.beta. to 4-.alpha. in the phorbol series, and indeed some of the diterpene skeleton structures carry hydroxy groups that must be esterified in order for inflammatory activity to be observed. However, these inactive compounds are quite few in number among the known phorboids, and no therapeutic utility has been demonstrated for them.
The phorbol esters, indolactams, polyacetates, diaminobenzyl alcohols, and bryostatins are generally found in plants, molds, and algae, or are synthetic in origin. Although they are found in many parts of the world, normal human contact with them is thought to be low. In contrast, the diacylglycerols are part of the functioning of virtually every type of animal cell and, thus, the undesirable activation of protein kinase C by the diacylglycerols may have a very widespread role in human diseases.
Thus, compounds capable of blocking the activation of, or inhibiting, protein kinase C by acting as specific pharmacological antagonists of the diacylglycerols at the diacylglycerol binding site on protein kinase C, would be valuable agents in the prevention and treatment of a wide variety of diseases in animals and humans. For example, the need for, and potential utility of, protein kinase C inhibitors/antagonists as agents for the treatment of cancer has received much attention [D. Corda, et al., Trends in Pharmacological Sciences 11: 471-473 (1990); G. Powis, Trends in Pharmacological Sciences 12: 188-194 (1991); S. Gandy and P. Greengard, Trends in Pharmacological Sciences 13: 108-113 (1992); B. Henderson and S. Blake, Trends in Pharmacological Sciences 13: 145-152 (1992)].
Protein kinase C comprises a family of eight or more closely related protein molecules [Parker, P. J. et al., Mol. Cell. Endocrin. 65: 1-11 (1989)]. Because of their high degree of relatedness they are referred to as "isozymes", "isotypes" or "isoforms". Occasionally the term "subtypes" is used, but this term is usually reserved to designate, as a subdivision, two or more variants of a single isotype.
The known isotypes of protein kinase C are: .alpha., .beta..sub.1, .beta..sub.2 and .gamma. (the "A-group"); .delta., .epsilon., .epsilon.' [Ono, Y. et al., J. Biol. Chem. 263: 6927-6932 (1988)], protein kinase C-L [Bacher, N. et al., Mol. Cell. Biol. 11: 126-133 (1991)], also known as protein kinase C .eta. [Osada, S. et al., J. Biol. Chem. 265: 22434-22440 (1990)], and .theta. [Osada, S.-I. et al., Mol. Cell. Biol. 12: 3930-3938 (1992) (the "B-group"); .zeta. and .iota. [Selbie, L. A. et al., J. Biol. Chem. 268: 24296-24302 (1993), also known as PKC.lambda. [Akimoto, K. et al., J. Biol. Chem. 269: 12677-12683 (1994)] (the "C-group"); and, .mu. [Johannes, F. -J. et al., J. Biol. Chem. 269: 6140-6148 (1994)] and PKD [Valverde, A. M. et al., Proc. Natl. Acad. Sci USA 91: 8572-8576 (1994) (the "D-group"). Members of the A-group require calcium ions for maximal activation, whereas the B-, C- and D-group members are thought to be largely calcium-independent for activation. The genes for each of the isotypes above have been cloned from one or more animal and yeast species and the clones have been sequenced; the relatedness of the genes and their product polypeptides is thus well established.
It is possible that the different protein kinase C isozymes have different biological roles, and published evidence supports this idea [Homan, E., Jensen, D. and Sando, J., J. Biol. Chem. 266: 5676-5681 (1991); Gusovsky, F. and Gutkind, S., Mol. Pharm. 39: 124-129 (1991); Borner, C., "The Role of protein kinase C in Growth Control", Sixth International Symposium on Cellular Endocrinology, W. Alton Jones Cell Science Center, Lake Placid, N.Y., Aug. 12-15, 1990; Naor, Z. et al., Proc. Natl. Acad. Sci. USA 6: 4501-4504 (1989); Godson, C., Weiss, B. and Insel, P., J. Biol. Chem. 265: 8369-8372 (1990); Melloni, E. et al., Proc. Natl. Acad. Sci. USA 87: 4417-4420 (1990); Koretzky, G. et al., J. Immunology 143: 1692-1695 (1989)]. For example, the stimulation of one protein kinase C isotype or a limited subset of protein kinase C isotypes might lead to undesirable results such as the development of inflammation [Ohuchi, K. et al., Biochim. Biophys. Acta 925: 156-163 (1987)], the promotion of tumor formation [Slaga, T., Envir. Health Perspec. 50: 3-14 (1983)] or an increased rate of viral replication in cells (i.e., de novo infection of cells and/or expression, assembly and release of new viral particles) [Harada, S. et al., Virology 154: 249-258 (1986)].
On the other hand, other protein kinase C isozymes might be responsible for the many beneficial effects observed when protein kinase C is stimulated by known protein kinase C activators in a variety of biological settings; such beneficial effects include the cessation of division of leukemic cells [Rovera, G., O'Brien, T. and Diamond, L., Science 204: 868-870 (1979)], multiplication of colonies of lymphocytes [Rosenstreich, D. and Mizel, S., J. Immunol. 123: 1749-1754 (1979)] and leucocytes [Skinnider, L. and McAskill, J., Exp. Hematol. 8: 477-483 (1980)] or the secretion of useful bioregulatory factors such as interferon-c [Braude, I., U.S. Pat. No. 4,376,822] and interleukin-2 [Gillis, S., U.S. Pat. No. 4,401,756].
Recent publications indicate that diacylglycerol binding sites exist on newly-described proteins which lack the kinase domain, and thus lack the kinase activity, of protein kinase C. One such protein is n-chimaerin, found in human brain [Ahmed et al., Biochem. J. 272: 767-773 (1990)] and the other is the unc-13 gene product of the nematode Caenorhabditis elegans, [Maruyama, I. and Brenner, S., Proc. Natl. Acad. Sci. USA 88: 5729-5733 (1991)]. The presence of the diacylglycerol binding sites on these two proteins was demonstrated by standard binding experiments with [.sup.3 H]phorbol 12,13-dibutyrate. These new proteins may have other enzymatic or biological activities which can be modulated by compounds which bind to their diacylglycerol binding sites. Thus, such compounds may have utility on non-protein kinase C biological targets.
Given that there are now numerous distinct biological entities bearing diacylglycerol binding sites, it would be highly desirable to obtain chemical compounds which could specifically and selectively target one or another type of diacylglycerol binding site, thus permitting one to selectively activate or inhibit one such site without affecting the others. Such compounds would be valuable experimental tools for studying the role of individual types of proteins bearing diacylglycerol binding sites as well as providing novel means for treating diseases in which protein kinase C or other diacylglycerol binding site-bearing proteins are involved.
There are several published reports describing chemical compounds capable of selectively distinguishing several diacylglycerol/phorboid-type binding sites in mouse skin [Dunn and Blumberg, op. cit. ] and in purified preparations of protein kinase C isotypes [Ryves, W. J., etal., FEBS Letters 288: 5-9(1991)]. However, in these studies, even the compounds showing the clearest differences in affinity for these distinct classes, namely phorbol 12,13-dibutyrate, 12-deoxyphorbol 13-isobutyrate, 12-deoxyphorbol 13-phenylacetate and thymeleatoxin, are only selective by a factor of 10-1000 in dissociation constant among the different binding sites. Furthermore, these compounds have potent skin inflammatory activity and are not desirable in human or animal medicine because of this toxicity.
Thus, to briefly recapitulate, two kinds of new compounds relating to diacylglycerol binding sites would be highly desirable. The first type would be capable of selectively activating one or a few useful, but not other, deleterious, diacylglycerol target sites. The second type would be capable of inhibiting, or antagonizing the stimulation of, one or more deleterious diacylglycerol binding site-bearing entities without blocking the useful ones. These kinds of compounds would be valuable agents for the study of diacylglycerol binding site-bearing entities and for the prevention or treatment of a wide range of human and animal diseases thought to involve protein kinase C or other entities under the control of diacylglycerol binding sites.
Earlier efforts to use the previously known phorboids themselves or to modify the structures of these known phorboids, have generally not been successful in producing useful compounds with toxicity low enough for use in humans.
It has been known for some time that several of the toxic, inflammatory and tumor-promoting compounds such as phorbol 12-tigliate 13-decanoate, mezerein, lyngbyatoxin and aplysiatoxin have anti-leukemic activity in mouse model tests [T. Sugimura, op cit., S. M. Kupchan and R. L. Baxter, Science 187: 652-653 (1975); S. M. Kupchan, et al., Science 191: 571-572 (1976); M. C. Territo and H. P. Koeffler, Br. J. Haematol. 47, 479-483 (1981)]. However, these compounds are all extremely toxic and are cancer suspect agents, thus eliminating them from consideration as human therapeutic agents.
Ganong, et al. [Proc. Nat. Acad. Sci. USA 83: 1184-1188 (1986)] tested a series of diacylglycerols and found no antagonistic activity in that series against the standard agonist, 1,2-dioctanoylglycerol. It is of particular note that several compounds tested in this work were modified in the hydroxymethyl portion of the diacylglycerol molecule, and these modifications produced only a loss of activity or a weakened activity that was not distinguishable from the agonist activity of 1,2-dioctanoylglycerol itself, a compound which is toxic to mouse skin [R. Smart, et al., Carcinogenesis 7: 1865-1870 (1986); A. Verma, Cancer Res. 48: 2168-2173 (1988)]. These hydroxymethyl-modified compounds were not antagonists in these tests and no utility was found. Similarly, Thielmann and Hecker [Forsch. Krebsforsch. Vol. VII, pp. 171-179 (1969), New York: Schattauer] found only a complete loss of biological activity in their study when the hydroxy group of the hydroxymethyl on phorbol 12,13-didecanoate was replaced with hydrogen or chlorine. Schmidt and Hecker [H. Lettre and G. Wagner (eds.), Aktuelle Probleme aus dem Gebiet der Cancerologie, Vol. III, 3rd Heidelberg Symposium, pp. 98-108. Berlin: Springer Verlag, 1971] also found that oxidation of the hydroxymethyl of phorbol 12,13-didecanoate to a carboxylic acid caused complete loss of activity in the assays used.
The hydroxymethyl group of the known phorboids (see structures above) has been thought to be required for biological activity, as detailed by Hecker (Hecker, E., Carcinogenesis, Vol. 2, eds. Slaga, Sivak and Boutwell, Raven Press, New York, 1978, pp. 11-48 and references cited therein). Indeed, it is stated therein that the replacement of the 20-hydroxyl in a phorbol ester "results in complete loss of biological activity". In another study, replacement of the hydroxy group of the hydroxymethyl (located at carbon 14) by chlorine or hydrogen in indolactam V gave rise to compounds with agonist activity weaker than but otherwise not distinguished from the agonist activity of the very toxic teleocidin class of tumor promoters [Irie et al., Int. J. Cancer 36: 485-488 (1985)]. Thus no utility beyond that of the toxic, hydroxymethyl-bearing parent indolactam-type compounds was found.
Schmidt and Hecker (Carcinogenesis, Vol. 7, ed. by E. Hecker et al., Raven Press, New York, 1982, pp. 57-63) studied the abilities of a series of diterpene phorboids to inhibit tumor promotion by the standard phorboid agonist tumor promoter phorbol 12-myristate 13-acetate (PMA). They found that, at low doses, some short-chain ester derivatives of phorbol were able to block the tumor promotion by PMA. However, all of the compounds that were active as antagonists at low doses are also very efficacious skin irritants themselves at slightly higher doses and most of them are also known to have tumor promoting activity. Thus, these short-chain esters still have toxic inflammatory and tumor promoting activity at doses only slightly different from those which would be needed to exhibit a therapeutic effect in mice.